The major histocompatibility complex (MHC) influences sexual selection in various vertebrates. Recently, MHC-linked social signaling was also shown to influence female fertility in horses (Equus caballus) diagnosed 17 days after fertilization. However, it remained unclear at which stage the pregnancy was terminated. Here we test if MHC-linked cryptic female choice in horses happens during the first days of pregnancy, i.e., until shortly after embryonic entrance into the uterus and before fixation in the endometrium. We exposed estrous mares to one of several unrelated stallions, instrumentally inseminated them with semen of another stallion, and flushed the uterus 8 days later to test for the presence of embryos. In total 68 embryos could be collected from 97 experimental trials. This success rate of 70.1% was significantly different from the mean pregnancy rate of 45.7% observed 17 days after fertilization using the same experimental protocol but without embryo flushing. Embryo recovery rate was not significantly dependent on whether the mares had been socially exposed to an MHc-dissimilar or an MHc-similar stallion. these observations suggest that MHc-linked maternal strategies affect embryo survival mainly (or only) during the time of fixation in the uterus. The major histocompatibility complex (MHC) plays a key role in the adaptive immune response of vertebrates 1. MHC molecules present antigen peptides to T cells enabling the immune system to recognize pathogens 2. The MHC also influences social signaling, as demonstrated in more than 20 species so far 1,3,4. In humans, for example, at least 16 studies provide evidence for MHC-related odors or odor preference 5 , and further studies report evidence for MHC effects on mate choice or sexual responsivity in our species 6. MHC-linked social signaling is also well established in the horse (Equus caballus) 7-10. MHC-linked social signaling may not only influence kin recognition and mate choice but also maternal investment into a pregnancy 1,3,4 , either to avoid male infanticide ("Bruce effect" 11,12) or as a late form of sexual selection that favors certain MHC genotypes over others 13. In the former case, MHC-linked signals that reveal the presence of a new dominant male can induce a pregnancy block 14,15 while MHC matching between male and female would not be expected to matter 16. In the latter case, i.e. if maternal investment into a given pregnancy reveals female preferences for certain male genotypes, higher rates of failed pregnancies would be predicted for MHC-similar than for MHC-dissimilar pairs 3,4. Indeed, human couples that share MHC antigens frequently suffer from a higher prevalence of recurrent spontaneous abortions than couples that do not share MHC antigens 17,18 , and Ober et al. 19 found reduced fecundity of MHC-similar couples within Hutterites, a group of people with usually large family sizes. These MHC effects seemed not linked to inbreeding 19. Women (or females in general) are not expected to have conscious control over these physiologic...